旅游多主题情感词典的构建方法  

Multi-theme Sentiment Lexicon Construction Method for Tourism

作　　者：林振宇 解吉波 杨腾飞 刘战 赵静 LIN Zhen-yu;XIE Ji-bo;YANG Teng-fei;LIU Zhan;ZHAO Jing(Hainan Provincial Key Laboratory of Earth Observation,Sanya 572000;Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094;East Sea Forecast Center of State Oceanic Administration,Shanghai 200136;School of Surveying and Land Information Engineering,Henan Polytechnic University,Jiaozuo 454000;University of Chinese Academy of Sciences,Beijing 100049,China)

机构地区：[1]海南省地球观测重点实验室,海南三亚572000 [2]中国科学院空天信息创新研究院,北京100094 [3]国家海洋局东海预报中心,上海200136 [4]河南理工大学测绘与国土信息工程学院,河南焦作454000 [5]中国科学院大学,北京100049

出　　处：《地理与地理信息科学》2021年第4期22-27,98,共7页Geography and Geo-Information Science

基　　金：海南省重大科技计划项目(ZDKJ2019006)。

摘　　要：面向旅游社交媒体大数据的分析和挖掘,该文提出一种旅游领域多主题情感词典的构建方法:首先,通过自然语言处理、机器学习,快速提取景区评价数据中旅游主题及其情感倾向,然后基于旅游多主题字典定义覆盖景区9类主题及细粒度种子主题词,最后针对景区情感倾向分析问题,根据词共现模型筛选与主题词典构成搭配的情感词,构建面向旅游领域的多主题情感词典。以海南省A级景区为例,基于上述构建的旅游领域多主题情感词典和景区网络关注度计算结果,对游客评论文本进行信息挖掘,并进一步结合GIS时空数据挖掘及网络关注度分析等方法,分析研究区域旅游景区游览信息的时空分布特征。结果表明,该文方法能有效监测景区各项主题的好评程度时空变化,验证了该方法的实用性和有效性。Facing the analysis and mining of big data of tourism social media,this paper proposed a method of constructing multi-theme sentiment lexicon.Firstly,the natural language processing and machine learning were used to extract the tourism themes and sentiment tendency from the evaluation data of scenic spots.Then nine kinds of themes and seed keywords were defined based on the multi-theme lexicon of tourism.Finally,aiming at the sentiment tendency analysis of scenic spots,the word co-occurrence model was used to select the sentiment words matched with the theme lexicon,and then build the multi-theme sentiment lexicon.A-level scenic spots of Hainan Province were used to verify the proposed method and lexicon.Based on the multi-theme sentiment lexicon and scenic spot attention index,the tourist comments were mined from tourism social media.The spatio-temporal method of GIS and network attention analysis were used to discover the spatial and temporal characteristics of tourism information of scenic spots.And the results showed that the proposed method can effectively monitor the temporal and spatial changes of the praise degree of each theme for scenic spots in detail.

关 键 词：旅游 多主题 情感词典 时空数据挖掘 自然语言处理 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]